How We Score NZ Investment Properties in 2026

A property investment score is a modelled rank of one listing against the same rules as every other listing in the set. It is not a registered valuation, a LIM, or a recommendation to buy. FindMyProperty publishes that rank so an active value-add investor or a yield-focused landlord can decide whether a TradeMe listing is worth an hour of due diligence.

This page is the 25 August 2026 refresh of the original February methodology note. The URL is unchanged. The numbers below match the scoring engine and Settings defaults — including verdict bands that the old copy had wrong (it quoted 80 / 65 / 45). If Settings are changed in production, the live score follows Settings, not this article.

Key takeaways for NZ investors

• Composite = 45% ROI + 17.5% timeline + 17.5% confidence + 20% location.
• Strong Buy ≥75, Buy ≥55, Maybe ≥35, Pass below 35 (August 2026 defaults).
• Flip vs Rental uses 15% modelled flip ROI and 9% gross yield. Both viable does not create a Hold label — the engine picks one path.
• Reno bands are COSMETIC $500/m², MODERATE $1,200, MAJOR $2,000, FULL_GUT $3,500, plus line items and 15% contingency.
• Some listings never get a score (new/renovated skip, extreme-metric hold). Override reno in the scenario modeller before you bid.

The four-component scoring model

Each component is 0–100. The composite is the weighted average. ROI is the largest weight because a beautiful kitchen in a 4% yield town still fails the job: return capital or pay a yield.

ROI score (45%)

If the recommended strategy is Flip, the ROI score is modelled flip ROI divided by 30%, times 100, capped at 100. If the strategy is Rental, it is gross yield divided by 15%, times 100, capped at 100. If the strategy is Pass, the engine takes the better of those two mappings. That is why a 15% flip can look “only mid-pack” on the ROI component: 15 / 30 = 50 on that axis before other components move the composite.

Default *strategy* targets remain 15% flip ROI and 9% gross yield. Those gates decide Flip vs Rental vs Pass. They are not the 100-point of the ROI component. Flip ROI itself is ARV minus purchase, reno, and holding, over total invested — the same shape as flip or hold in August 2026.

Timeline score (17.5%)

Eight weeks or fewer scores 100. From week 9 to 16 the score drops five points per extra week. Beyond 16 weeks it drops three points per week from a lower base. A long structural job can still rank if ROI and location are strong; it will not look like an eight-week cosmetic do-up.

Confidence score (17.5%)

Confidence is the average of ARV confidence (comparable sales) and a rental cap from bond sample count: each Tenancy Services sample adds about five points, capped at 100. Sparse comps or a thin bond book lower the composite even when the yield print looks aggressive. That is deliberate. A number without evidence should not outrank a slightly worse number with a stack of bonds.

Location score (20%)

Location starts from projected territorial-authority population growth, then weights nearby amenity: employment hubs 40%, daily convenience 25%, education access 20%, residual growth 15%. It is not a suburb “vibe” score. Pair it with suburb intelligence when you care about the street, not the TA average.

Vision, renovation dollars, and what we skip

Before financials, listing photos go through vision. The model grades roof, exterior, interior, kitchen, bathroom, and visible structural issues, then assigns a renovation level. A cheaper text signal from the listing write-up runs first so we do not spend vision on copy that is obviously not a do-up. The cost model then applies a per-square-metre band to floor area, adds flagged line items (re-roof, moisture, Healthy Homes allowance when the copy supports it), and a 15% contingency on that total.

These are August 2026 internal bands in the renovation cost module — not a builder quote. The older blog quoted $400–$1,500 “light / moderate / heavy” bands that the engine no longer uses. Walk through photo flags in how the renovation cost estimator works. Scope that actually pays is bang-for-buck renovations.

Not every passed listing is scored. If the marketing reads new-build / fully renovated / move-in ready and vision agrees the programme is negligible (about $7,500 or less and a couple of weeks), we omit a composite so turnkey stock does not crowd value-add ranks. If modelled flip ROI or yield prints in an extreme band, publication can wait on manual review. Pipeline shape is summarised on methodology.

Flip vs rental — there is no Hold label

After flip and rental models run, strategy logic compares modelled flip ROI to 15% and gross yield to 9% (moderate risk default). If only flip clears, the label is Flip. If only yield clears, the label is Rental. If neither clears, Pass. If both clear, the engine does **not** output Hold: it picks Flip when flip ROI is more than 1.5× yield and the market path still favours a flip; otherwise Rental. You can still buy a Flip-labelled listing to hold — that is your exit, not the tag on the card. Work that decision in flip or hold.

Verdict bands (August 2026 defaults)

Verdicts are cut from the composite, with one extra Pass path: if strategy is Pass and the composite is under 40, verdict is Pass even if the Maybe floor would otherwise apply. Frontend badges use the same 75 / 55 / 35 cutovers.

“The score is a starting point, not a final answer. Verify renovation costs with a local builder before you commit to a deal.”

What the score cannot see

Vision cannot see behind linings. ARV cannot see an unconsented sleepout. Bond medians cannot see the tenant you will actually get. We do not score flood or meth from photos unless the image itself shows damage, and we do not sell insurance. Cover gaps after you buy are landlord insurance for NZ investors. The human checklist is how to analyse an investment property before you buy.

• Override reno in the scenario modeller if the photos look heavier than the band.
• Treat 15% flip / 9% yield as strategy gates, not as “the deal is safe.”
• If confidence is low, get more comps or a valuer before you stretch price.
• FindMyProperty does not score commercial listings or quote cover.